Research on Default Risk Prediction of Listed Companies’ Green Credit Based on Deep Learning Algorithm

The prediction of green credit default on deep learning algorithm is to predict the in the green credit portfolio of enterprises. We use recursive feature elimination (RFE) algorithm to train the DL model, and then apply it to predict the default risk of green credit of listed companies. The RFE algorithm iteratively deletes features from the input data until the performance on the verification set cannot be improved. In our research, we found that the combination of RFE algorithm and logical regression as outlier detector can improve the prediction performance by 2%, and the prediction probability as the input of the risk control system can be used to evaluate the effectiveness of the management’s decision-making ability.